遥感及再分析降水产品在缺资料干旱内陆盆地的适用性评估

山区降水较集中，但降水测站多位于山谷或人口密集区，代表性差。遥感和再分析降水产品能提供时空分布连续的数据，不受地形条件限制。柴达木盆地中心属干旱荒漠区，水是制约该区开发的首要条件，其四周属高寒山区，降水相对较多，但降水监测十分薄弱。为获取该区相对精确的降水时空分布信息，本文评估了4套高分辨率降水产品（CMADS、TRMM、GPM和MSWEP）的适用性。首先基于地面站点数据评估它们在不同时空尺度上的精度，并分析它们在柴达木盆地的空间分布和年内分配特征。然后，以盆地东南隅的无测站山区香日德河流域为研究区，利用降水产品驱动SWAT模型来评估它们的分布式水文模拟适用性。结果表明：(1) MSWEP在年、月尺度上与站点降水的吻合程度最高（R≥0.79,PBIAS=0.5%），其次是GPM和TRMM,CMADS精度最低（R≥0.64,PBIAS=5.8%）；(2)从降水精度与站点高程的关系来看，降水产品在相对低海拔区容易高估站点降水，而在相对高海拔区常低估实际降水；(3)在香日德河流域，MSWEP(NSE=0.64)在基准期（2009—2012年）的径流模拟表现明显好于其它降水产品（NSE=0.3...

Applicability Analysis of Satellite-based and Reanalysis Precipitation Products in Poorly-Gauged Arid Inland Basins

Precipitation is often concentrated in mountainous regions, but rain gauges mostly cluster in flat valleys or populated regions and thus have poor spatial and temporal consistency. Satellite-based and reanalysis precipitation products can provide precipitation data without being restricted by local terrain condition. However, these products are usually subject to systematic errors, which may vary with time and topographic condition. Thus, it is essential to evaluate and analyze the regional applicability of different precipitation products. The Qaidam Basin is one of the typical alpine arid inland basins in China. Its center part belongs to an arid desert region, and water scarcity hampers the socioeconomic development of this region. The Qaidam basin is surrounded by mountainous regions which are the “water tower” in the world with concentrated precipitation but sparse rain gauge networks. The lack of precise precipitation observation in space and time in mountainous regions can result in unreliable hydrologic modeling and unsustainable water management. To obtain accurate precipitation data, this study collected and evaluated four widely used precipitation products with higher spatiotemporal resolution, including CMADS, TRMM, GPM, and MSWEP. First, we evaluated the suitability of these four precipitation products in the Qaidam Basin against gauge-based precipitation data, using monthly precipitation data from nine meteorological stations and 13 hydrological stations. The spatiotemporal characteristics and the site level errors of the four precipitation products were also analyzed. Then, we concentrated on the typically poorly gauged basin, Xiangride River Basin located in the southeastern Qaidam Basin. The Four precipitation products were input to the semi-distributed hydrological model, Soil and Water Assessment Tool (SWAT), to evaluate the simulated streamflow by comparison with observed streamflow from the gauge data at the watershed outlet (Qianwaebo hydrological station). Results show that: (1) At yearly and monthly scales, MSWEP achieved the highest agreement (R ≥ 0.79, PBIAS = 0.5%) against the observed data, followed by GPM and TRMM, while CMADS had the worst performance (R ≥ 0.64, PBIAS = 5.8%); (2) All precipitation products generally overestimated precipitation at lower elevations and underestimated precipitation at higher elevations; (3) At the basin scale, MSWEP performed best (NSE = 0.64 vs. 0.36~0.59 for other datasets) in the baseline period (2009—2012), while CMADS performed best (NSE = 0.75 vs. 0.53~0.68 for other datasets) in the impacted period (2013—2016). This study can provide important support for obtaining accurate precipitation information and improving water resources management in poorly gauged arid mountainous region.